Stock Market crashes Zhiguo He Professor of Finance University of Chicago Booth School of Business and NBER Joint with Jiangze Bian Kelly Shue and Hao Zhou Excessive Leverage and fire sales are considered to be the underlying mechanism of many crises in financial markets ID: 742232
Download Presentation The PPT/PDF document "Leverage-Induced Fire Sales and" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Leverage-Induced Fire Sales and Stock Market crashes
Zhiguo HeProfessor of FinanceUniversity of Chicago, Booth School of Business; and NBERJoint with Jiangze Bian, Kelly Shue, and Hao ZhouSlide2
Excessive Leverage and fire sales are considered to be the underlying mechanism of many crises in financial markets
1929 US stock market crash
2007/08
financial and housing market crisesChinese stock market crash in 2015 Yet, limited empirical evidence on fire-sale, and not in the context of leverage Coval and Stafford (2007) and Edmans, Goldstein and Jiang (2012): fire-sale of mutual funds due to fund outflows Ellul, Jotikasthira, and Lundblad (2011): fire-sale of downgraded corporate bonds due to regulatory constraintCampbell et all (2011, foreclosed housing); Pulvino (1998, commercial aircraft)This paper: direct evidence of leverage-induced fire sales Account-level trading data for margin accounts in Chinese stock market in 2015
IntroductionSlide3
Background (1)
Chinese stock market rises quickly in the first half of 2015 and crashed thereafter
Shanghai
C
omposite Index: started around 3100 on Jan 2015, peaked 5166 on
J
une 15
th
, 2015, then collapsed to
3663 at the end of July
Two kinds of margin accounts were popular starting mid-2014 in Chinese stock market
Brokerage-financed margin system
Similar to US margin trading (initial margin, maintenance margin,
etc
)
Tightly regulated,
with
minimum
initial margin and maintenance margin
Shadow-financed margin system
Unregulated grey area; higher leverage, lower maintenance margin, and large cross-sectional variations
Regulation tightening is expected to come…Slide4
Background (2)
forced fire-sale of leveraged
accounts
is accused as the leading cause of china’s stock market crashMay 22 2015, CSRC (China Securities Regulation Commission) announces to start investigating “illegal” shadow-financed margin accountsJune 12 2015, release draft rules “ban on shadow-financed margin accounts”What do data tell us?Slide5
Data description
detailed account level daily trading records during crisis (may-July 2015)
Brokerage-financed margin accounts from a leading brokerage in China, with a market share of ~10% in brokerage margin credit
Shadow-financed margin accounts from
a
leading web-based peer-to-peer lending platform
H
ard to estimate its market share in shadow margin accounts; one reasonable estimate is about 5% (in terms of debt)
each individual account in both categories:
Daily stock holdings and trading
Daily asset and debt data, hence leverage defined as asset/(asset-debt)
Account maximum allowable leverage (
P
ingcang
Line,
平仓线
)
Stock daily information: prices, returns, outstanding shares,
etc
Slide6
Leverages and Market index (1)
Leverage: Asset/Equity. Asset-weighted and equity-weighted are quite different!
index, right scale
leverage, left scaleSlide7
Leverages and Market index (2)
Leverage: Asset/Equity. Leverage dispersion rises dramatically right after market crash
index, right scale
leverage, left scaleSlide8
Leverage induced fire-sale
:Account level evidence (2)
: the maximum allowable leverage of this account
So-called Pingcang Line possible: cannot sell if stocks hit -10% daily limit rule; lenders are unsophisticated investors as well Define Proximity to Pingcang Line
Sort accounts into equally-spaced bins by
Slide9
Leverage induced fire-sale
:Account level evidence (1)
Account-stock-date level regression:
Stock-date fixed effect and account fixed effect Identification comes from account ’s time-varying proximity
to its Pingcang
Line
Leverage-induced selling implies that
increases with
Slide10
Leverage induced fire-sale
:Account level evidence (2)
Benchmark: classify accounts with
as “fire-sale accounts,” cut-off rule Robustness later: using these ’s as weights Slide11
Leverage-induced selling on Brokerage and Shadow
Same regression, separately for Brokerage (higher leverage, mean 1.43) and Shadow (higher leverage, mean 6.61)Slide12
Leverage and Proximity to
Pingcang (1)
Proximity to
Pingcang line determines leverage-induced Fire-saleLeverage will amplify shocks…For accounts with the same Proximity, those with higher leverage should sell more aggressivelyBecause they will hit the Pingcang Line more likelyShadow sample, adding leverage dummy and interactionsSlide13
Leverage and Proximity to
Pingcang (2)Slide14
The role of Regulation tightening? (1)
New releases of regulations on shadow accounts on 5/22 and 6/12
5/22 initial announcement; 6/12, detailed draft
repeat the same exercises, for trading weeks before and after these two eventsFor both brokerage and shadow accountsSlide15
The role of Regulation tightening? (2)Slide16
Leverage induced fire-sale
:Stock level evidence (1)
If stock
is held by more fire-sale accounts, it will be sold more heavily by these accountsRun Regression Fire-sale accounts: accounts with at the beginning of
is stock ’s F
ire-Sale
E
xposure, defined as
Slide17
Leverage induced fire-sale
:Stock level evidence (2)
(1)
(2)(3)(4)VARIABLESNet selling of fire-sale accounts Fire Sale Exposure (FSE) 0.0996***
0.102***
0.
102
***
0.102
***
(
0.02
21
)
(
0.02
5
9
)
(
0.02
59
)
(
0.025
9
)
Return Volatility
X
X
Size
(Market Cap)
X
X
Turnover
X
X
Past
10-day cum.
return
X
X
Past
10-day daily return
X
Stock FE
X
X
X
Date FE
X
X
X
Observations
1
16,809
1
16,809
1
16,809
1
16,809
R-squared
0.1
44
0.1
86
0.1
86
0.18
7Slide18
Stock Returns following fire-sale (1)
do leverage-induced fire sales cause subsequent low stock return?
Empirical predictions of fire-sale:
Stocks with high underperform in the short-run but long-run reversalTwo methodsDouble sort on past return and ; long-short strategy based on Address the concern of momentum….. Regress stock returns on with various controls
Slide19
Stock Returns following fire-sale (2)
Double sort: Each day, we
First, sort stocks into quartiles by past-10-day returns;Second, sort each quintile into deciles by (recall this is measured at the beginning of date )Cumulative abnormal return of long-top-short-bottom decilesLeverage induced fire-sale storyNegative abnormal return of this long-short strategy, but disappears in long-run
Slide20
Stock Returns following fire-sale (3)Slide21
Stock Returns following fire-sale (4)
regression
Abnormal return is based on CAPM with stock beta calculated using 2014 data
and Model prediction for small but for large
Slide22
Stock Returns following fire-sale (5)
CAR identified by
FSE
Robust standard errors in parentheses, clustered at date levelControls: return volatility; market cap; past 10-day daily returns; past 10-day cumulative return; turnover; stock fixed effect; date fixed effect 1 Day3 Days5 Days10 Days20 Days40 DaysFSE-1.422***
-3.529***
-
5.089
***
-
5.
924
***
-
2.
827
***
0.2
9
0
SE
(
0.
312
)
(
0.5
94
)
(
0.
910
)
(
1.2
64
)
(
0.9
24
)
(
0.5
70
)Slide23
Brokerage vs ShadowSlide24
Brokerage vs Shadow
Fire-sale account cut-off
;
high leverage greater proximity Slide25
Shadow or brokerage?
Robust standard errors in parentheses, clustered at date level
1 Day3 Days5 Days10 Days20 Days40 Days
FSE of
shadow
-
2.
368
***
-
5.
903
***
-
8.
935
***
-
11.
65
***
-
3.
620
*
0.5
47
SE
(
0.
615
)
(
1.
339
)
(
1.
919
)
(
2.
404
)
(
1.9
37
)
(
0.8
44
)
FSE of brokerage
-
0.
621
***
-
1.
519
***
-
1.
829
**
-
1.052
-
2.
170
***
0.0
725
SE
(
0.2
22
)
(
0.
444
)
(
0.6
58
)
(
0.7
43
)
(
0.4
78
)
(
0.6
62
)Slide26
Robustness: Market up and down (1)
Leverage-induced fire-sale effect should exist in both up and down market conditions
Need to sell to keep the account safe
positive feedback of leverage spiral stronger fire-sale effect in market downturn Slide27
Robustness: Market up and down (2)Slide28
Robustness: Hitting price limit (1)
Chinese stock market sets a daily price limit for each stock
Sock return cannot exceed [-10%,10%] range
More fire-Sale of this stock if other stocks in this account hit their 90% price limitsFor stock in account , construct “fraction of hitting limit” = Slide29
Robustness: Hitting price limit (
2)Slide30
Robustness: constructing
based on weights
Constructing stock level fire-sale Exposure
based on : number of shares of stock in account
Numerator: weighted sum of shares of stock in account
; if account
belongs to group
then the weight is
Again, leverage is measured at the beginning of date t
Slide31
Direct evidence on Leverage-induced fire sales
The closer to the maximum allowable leverage, the more you sell (including both forced sale and preemptive sale)
The resulting fire-sale leads to negative abnormal return in the short-run
Regulated brokerage vs unregulated shadow margin accountsBrokerage margin accounts are dominant in holdings, but relatively low fire-sale pressureShadow margin accounts are the major force of leverage-induced fire-sale in 2015 stock market crashBian et al (2017) study the amplification effect through the lens of a network frameworkFull-blown amplification and propagation requires a structural model, work to be done in the futureConcluding RemarksSlide32
Leverages and Market index
Leverage: Asset/Equity. Weighted by equityUnregulated shadow has higher leverage
index, right scale
leverage, left scaleSlide33
Brokerage Leverages
index, right scale
leverage
, left scaleSlide34
Shadow Leverages